As well as, Cisco and Nvidia will put money into cross-portfolio expertise to deal with widespread challenges like congestion administration and cargo balancing, making certain that enterprises can speed up their AI deployments, Patel acknowledged.
The distributors mentioned they might additionally collaborate to create and validate Nvidia Cloud Companion (NCP) and Enterprise Reference Architectures primarily based on Nvidia Spectrum-X with Cisco Silicon One, Hyperfabric, Nexus, UCS Compute, Optics, and different Cisco applied sciences.
Historical past of Cisco, Nvidia collaborations
The announcement is simply the most recent enlargement of the Cisco/Nvidia partnership. The businesses have already labored collectively to make Nvidia’s Tensor Core GPUs accessible in Cisco’s Unified Computing System (UCS) rack and blade servers, together with Cisco UCS X-Collection and UCS X-Collection Direct, to help AI and data-intensive workloads within the information heart and on the edge. The built-in bundle consists of Nvidia AI Enterprise software program, which options pretrained fashions and improvement instruments for production-ready AI.
Earlier this month, Cisco mentioned it has shipped the UCS C845A M8 Rack Server for enterprise information heart environments. The 8U rack server is constructed on Nvidia’s HGX platform and designed to ship the accelerated compute capabilities wanted for AI workloads reminiscent of LLM coaching, mannequin fine-tuning, massive mannequin inferencing, and retrieval-augmented technology (RAG).
The businesses are additionally collaborating on AI Pods, that are preconfigured, validated, and optimized infrastructure packages that prospects can plug into their information heart or edge environments as wanted.
The Pods are primarily based on Cisco Validated Design principals, which offer a blueprint for constructing dependable, scalable, and safe community infrastructures, in response to Cisco. The Pods embrace Nvidia AI Enterprise, which options pretrained fashions and improvement instruments for production-ready AI, and are managed via Cisco Intersight.
